PostgreSQL 19 Release Delayed by AI Bug Reports

PostgreSQL 19 is likely to be released in November rather than in the project's usual window between the middle of September and the middle of October, according to a blog post by Robert Haas published on Friday, October 30, 2026. The delay comes after several large features were reverted late in the beta period and after a surge in bug reports found with large language models.

Release timing and late reverts

Over the past decade, the timing of PostgreSQL major releases has been predictable, with releases arriving no earlier than the middle of September and no later than the middle of October. This year, PostgreSQL 19 most likely will not be released until November, and the project could decide to postpone it further. Between August 27 and September 15, several rather large features were reverted: UPDATE/DELETE FOR PORTION OF, SQL Property Graph Queries (SQL/PGQ), and ALTER TABLE ... MERGE PARTITION/SPLIT PARTITION. Major feature reverts are not especially uncommon immediately after feature freeze, which each year occurs on April 8, but it is unusual to see multiple major features pulled at such a late date.

LLM bug reports increase the workload

Robert Haas attributes the change to LLMs. Between April 8, 2025 and September 25, 2025, there were 1,037 commits to PostgreSQL's master branch and 724 commits to REL_18_STABLE. In 2026, during the same date range, there were 1,443 commits to the master branch and 1,169 commits to REL_19_STABLE. Commits to the master branch between feature freeze and September 25 declined slightly from last year to this year, while back-patched commits jumped by more than 60%. During 2025, there were 248 commits to REL_17_STABLE, and during 2026, there were 433 commits to REL_18_STABLE, meaning back-patched commits to at least one prior release jumped by more than 74% in a single year. A huge number of people on the Internet are now reporting bugs in PostgreSQL that they found using LLMs, and committers have been very busy fixing them.

Quality, review time, and long-term questions

Two factors are at play. First, because many more bugs are being found, fixes take more time, and LLM-drafted fixes are not always correct or compliant with PostgreSQL standards, so they still need review and commitment by humans. Second, time that would have been spent reviewing release changes was diverted into addressing security issues and other bugs in back branches, so it took longer to determine what needed to be reverted before shipping version 19. Haas says the delay is not particularly concerning and emerges directly from the project's concern for quality. The long-term picture is less clear: with forty years of bugs that humans were unable to find but LLMs can discover, he is unsure how PostgreSQL will catch up and continue to innovate. He maintains a high estimation of the code base, says it is not perfect, and argues that LLMs will require developers to raise their game as every feature is scrutinized by AI.

Sources: Lobsters